# Tag Archives: pre-election polls

## 2016 Pre-election Model – Calculating the Expected Electoral Vote

Richard Charnin
Aug. 29, 2017

2016 Pre-election Model – Calculating the Expected Electoral Vote

This is for those interested in Electoral Vote math based on pre-election polls. It discusses basic probability and spreadsheet functions. You won’t see a discussion of this anywhere else.The MSM doesn’t care for critical thinking. Perhaps because they are incapable of it.

One of the methods I have used in pre-election forecast modeling is to calculate the Expected Recorded Electoral Vote as well as the True Vote. Important Note: the RECORDED EV is based on MSM pre-election polls which are usually biased for the establishment candidate. In 2016, Clinton was the establishment candidate.

As I did not have 51 state pre-election polls, I used the following method to estimate them based on the average of nine pre-election national polls and Party-ID:

1) Each state’s estimated Party-ID was calculated using the proportional change from the 2012 National Party-ID to the 2016 Gallup National Voter affiliation survey: 40% Independents, 32% Democrats and 28% Republicans.

2) The average vote shares of nine national pre-election polls were applied to the Party-ID of each state to derive the projected state vote shares.

The Expected EV is based on state win probabilities. Calculating the pre and post-election TRUE EV is much more complicated.

In the 2016 Forecast Model, Trump’s Expected EV (before undecided voters) was 305.5, exactly matching his recorded 306 EV. His Snapshot 307 EV is the sum of the EVs for states that he was projected to win. Trump led the weighted average pre-election polls (before undecided voter allocation) by 44.1-43.1%.

View the Recorded votes and two True Vote Models for all the states:

The following steps calculate the Expected RECORDED EV:
1. Using state forecasts derived from the National Gallup Voter Affiliation survey, calculate the probability P(i) of winning each state using Trump’s projected 2-party vote share. Assume a 3.0% margin of error.
P(i) = normdist(Trump%/(Trump%+Clinton%),0.5,.03/1.96,true)

2. Multiply the state win probability by the state electoral vote.
S(i) = P(i)* EV(i), i =1,51
3. Expected EV = sum [P(i)* EV(i)], i = 1,51

State Electoral votes are in the range B129:B179
Trump’s state forecasts are in the range D129:D179
Corresponding state win probabilities are in the range J129:J179

The Expected EV calculation is in cell I128.
Expected EV = 305.5 = sumproduct(J129:J179, B129:B179)

Posted by on August 29, 2017 in 2016 election, electoral vote

## MY COMMENTS TO THE MSM ON THE RIGGING OF THE 2016 PRE-ELECTION POLLS

The MSM just interviewed the authors of  Shattered: Inside Hillary Clinton’s Doomed Campaign on the reasons for Clinton’s loss.  I commented to Chris Mathews and Brian Williams of MSNBC as well as FOX and CBS on how MSM pollsters rigged the pre-election polls for Clinton.

FYI: Your guests may not have looked at my 2016 Election model. It was based adjustments to final pre-election polls which were biased for Clinton. The Democratic Party-ID share was overstated at the expense of Independents who went solidly for Trump. In addition, there is strong evidence that votes were stolen from Jill Stein – by Clinton.

The 2016 Model projected Trump’s 306 RECORDED EV. But he actually had approximately 351 TRUE EV after adjusting for late undecided voters. https://richardcharnin.wordpress.com/2016/11/07/2016-election-model-forecast/

Recorded Vote: Clinton 48.3-46.2%, Trump 306-232 EV
Recorded Vote Forecast: Trump 44.4-42.9% with 306-232 EV
True Vote Model: Trump 48.5-44.3% with 351-187 EV

Note: I exactly forecast the RECORDED EV in the last three elections: 365, 332, 306. In each case the winner did better in the True Vote than the Recorded vote.

Posted by on April 24, 2017 in 2016 election

## Why the 2016 pre-election polls, unadjusted exit polls and recorded vote are all wrong

Richard Charnin
Dec.30, 2016

The 2016 election was different in kind from prior elections; the Democrat was the establishment candidate. It was established beyond a reasonable doubt that the primaries were stolen from Bernie Sanders by the DNC which colluded with the media.

Some analysts claim that the 2016 unadjusted state exit polls prove that the election was rigged for Trump. But just because the polls were excellent indicators of the True Vote in the past does not prove that they were accurate in 2016.

Are we supposed to believe that the MSM would not rig the unadjusted exit polls to match the rigged  pre-election polls  to make it appear that Clinton was the winner?

Exit pollsters at  Edison Research never reveal the location of precincts, votes and survey results. The only way to prove that the unadjusted exit polls are correct (and the published results bogus) is 1) to reveal the complete exit poll timeline and the data for all precincts polled and 2) a True Vote analysis based on historical and current independent data.

True Vote analysis indicates that Trump won the popular and electoral vote and that pre-election and exit polls were rigged for Clinton by inflating Democratic Party-ID. True Vote Models were based on a) national Gallup Party-ID voter affiliation and b) returning 2012 voters.

As usual, state and national exit polls were forced to match the recorded vote. This was the first election in which the media discussed election fraud – but avoided the obvious U.S. suspects from prior elections and the rigged voting machines, illegal and disenfranchised voters. Now that the MSM finally admits election fraud, they blame it on the Russians! And don’t report the proven fact that the primary was rigged for Clinton.

Party-ID
Nine Pre-election polls (average): 28.8 Ind – 38.7 Dem- 31.9 Rep.
Final National Exit Poll (CNN): 31 Ind – 36 Dem – 33 Rep.
Gallup national voter affiliation survey: 40 Ind -32 Dem -28 Rep. https://docs.google.com/spreadsheets/d/1R9Y3ae2uyW8SUxVUnnOt9ZyvheAxa0fAhesAw_nhciM/edit#gid=505041111

Nine Pre-election polls
Clinton won the average: 45.8-43.3%
Trump won the average Gallup-adjusted poll: 44.4-42.9%
Trump won Independents: 43.6-33.8%

Final  National Exit Poll (forced to match the Recorded Vote)
Clinton won the reported vote: 48.2-46.2%.
Clinton won the National Exit Poll: 47.7-46.2%.
Trump won Independents by just 46-42% – a 5.8% discrepancy from the pre-election polls which he led by 9.8%. This anomaly is additional evidence that Trump won the True Vote.

Clinton won the polls: 49.6-43.6%
Clinton won the corresponding recorded vote: 49.3-45.2%

States not exit polled
Trump won: 50.4-43.7%

True Vote
Trump led the True Vote Model (three scenarios of his share of late undecided voters)
– Scenario I:  47.5-45.1%, 306 EV (50% undecided)
– Scenario II: 47.9-44.7%, 321 EV (60% undecided)
– Scenario III: 48.3-44.3%, 351 EV (70% undecided)

The True Vote Model analysis based on a plausible number of returning voters from the prior election  confirmed the three scenarios: https://richardcharnin.wordpress.com/2017/04/29/university-of-virginia-study-20-of-trump-voters-were-former-obama-voters/

The National Election Pool of six media giants funds exit pollster Edison Research. The published results are always forced to match the recorded vote which implies zero election fraud. But there is always election fraud.  Historically, unadjusted state and national exit polls always favored the Democratic candidate, but there was  a RED shift from the Democrat in the poll to the Republican in the recorded vote.

The True Vote Model indicates that the 1988-2008 unadjusted exit polls were accurate.
https://richardcharnin.wordpress.com/2014/09/14/summary-2004-2012-election-forecast-1968-2012-true-vote-model/

1 Comment

Posted by on December 30, 2016 in 2016 election

## 2016 ELECTION MODEL -10/30 UPDATE – TRUMP SURGING

Richard Charnin
Oct. 30, 2016

The 2016 Election Model uses eight recent polls adjusted for party-ID weights and undecided voters.

Clinton leads the 8-poll average 45.4-42.5% with 309 EV. But when the over-weighted Democratic party-ID split is replaced by the Gallup party affiliation  survey and 75% of undecided voters are allocated to Trump, he leads by 48.2-44.1%  with a 336-202 EV lead..

The model indicates that Trump has a 92% popular vote win probability.

Note: Undecided voters typically break 75% for the challenger. Trump is the assumed challenger and HRC/Obama is the incumbent.

8-Poll…………………………….. Electoral Vote….. Trump Popular Vote
Average….. Clinton Trump….. Clinton.. Trump…Win Prob (3% MoE).
Poll………… 45.4%… 42.5%……. 309…… 229……..14%
Undec……… 44.1%… 48.2%…….202…… 336……..92%

% Ind ……… 32.5%… 45.1%

Party ID.. Ind…. Dem…. Rep
Avg Poll.. 27.5% 39.0% 31.9%

ABC/Washington Post

 ABC/WP Party-ID Stein Clinton Trump Johnson Ind 29% 1% 38% 58% 3% Dem 37% 1% 94% 5% 0% Rep 29% 0% 4% 91% 2% Total 95% 0.7% 47.0% 45.1% 1.5% Poll 97% 2.0% 46.0% 45.0% 4.0% Gallup Adj. 99.2% 0.7% 46.4% 50.3% 1.8%
 Gallup Adj. Party-ID Stein Clinton Trump Johnson Ind 40% 1% 38% 58% 3% Dem 32% 1% 94% 5% 0% Rep 28% 0% 4% 91% 2% Total 99.2% 0.7% 46.4% 50.3% 1.8% Votes 128,021 930 59,905 64,914 2,272 EVote 538 0 187 351 0

 Trump % Rep Trump 87.0% 89.0% 91.0% 93.0% 95.0% % Ind Trump 62% 50.8% 51.3% 51.9% 52.4% 53.0% 58% 49.2% 49.7% 50.3% 50.8% 51.4% 54% 47.6% 48.1% 48.7% 49.2% 49.8% Clinton 62% 45.9% 45.4% 44.8% 44.2% 43.7% 58% 47.5% 47.0% 46.4% 45.8% 45.3% 54% 49.1% 48.6% 48.0% 47.4% 46.9% Margin 62% 4.8% 6.0% 7.1% 8.2% 9.3% 58% 1.6% 2.8% 3.9% 5.0% 6.1% 54% -1.6% -0.4% 0.7% 1.8% 2.9% Vote Margin (000) 62% 6,196 7,630 9,064 10,498 11,932 58% 2,100 3,533 4,967 6,401 7,835 54% -1,997 -563 871 2,304 3,738

Posted by on October 30, 2016 in 2016 election

## A Tale of Two Pre-election Polls

Richard Charnin
Oct. 21, 2016

This analysis illustrates how polling results (and electoral votes) are manipulated due to the over-weighting of Democrats and under-weighting of Independents. The misleading poll results are compared to the more accurate Gallup party voter affiliation weighting.

The Quinnipiac poll has Clinton leading Trump 47-40%.
Given the internal poll shares, the Party-ID split is
Ind 26%, Dem 40%, Rep 34%.
Using the Four-way Election Model, Clinton wins by 444-94 Electoral Votes.

Assuming the Gallup party affiliation survey
(Ind 40%, Dem 32%, Rep 28%):
Clinton wins by 45.4-40.5% with 354-184 EV.

The Rasmussen poll has Trump leading Clinton 43-41%.
Given the internal poll shares, the Party-ID split is
Ind 32%, Dem 40%, Rep 28%.
Using the Four-way Election Model, Trump wins by 327-211 Electoral Votes

Assuming the  Gallup party affiliation survey
(Ind 40%, Dem 32%, Rep 28%):
Trump wins in a landslide by 45.4-36.5% with 496-42 EV.

 Quinnipiac Match poll Party-ID Pct Clinton Trump Johnson Stein Ind 26% 38% 42% 8% 12% Dem 40% 91% 4% 2% 3% Rep 34% 4% 80% 10% 6% Total 100% 47.64% 39.72% 6.28% 6.36% Votes 129,106 61,506 51,281 8,108 8,211 Elect Vote 538 444 94 0 0 Quinnipiac Match Gallup Party-ID Pct Clinton Trump Johnson Stein Ind 40% 38% 42% 8% 12% Dem 32% 91% 4% 2% 3% Rep 28% 4% 80% 10% 6% Total 100% 45.44% 40.48% 6.64% 7.44% Votes 129,106 58,666 52,262 8,573 9,605 Elect Vote 538 354 184 0 0
 Rasmussen Match poll 41 43 5 3 Party-ID Pct Clinton Trump Johnson Stein Ind 32% 22% 47% 18% 13% Dem 40% 77% 15% 3% 6% Rep 28% 11% 78% 8% 3% Total 100% 40.92% 42.88% 8.80% 7.40% Votes 129,106 52,830 55,360 11,361 9,554 Elect Vote 538 211 327 0 0
 Rasmussen Match Gallup Party-ID Pct Clinton Trump Johnson Stein Ind 40% 22% 47% 18% 13% Dem 32% 77% 15% 3% 6% Rep 28% 11% 78% 8% 3% Total 100% 36.52% 45.44% 10.08% 7.96% Votes 129,106 47,149 58,666 13,014 10,277 Elect Vote 538 42 496 0 0

Polling Data

http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton_vs_johnson_vs_stein-5952.html

Four-way 2016 Election Model

Posted by on October 21, 2016 in 2016 election

## Is the corporate media reporting Jill Stein’s true polling numbers?

Richard Charnin
Sept. 18, 2016

Just published: 77 Billion to One: 2016 Election Fraud

Is the corporate media reporting Jill Stein’s true polling numbers? It’s obvious that the corporate media does not want her in the debates. It would radically change the dynamic of the race.

Stein has just 3% in the polls and needed 15% to qualify for the debates. If Jill got in the debates, her visibility would skyrocket, her poll shares would increase and Hillary Clinton’s shares would decline..

How many Independent and Democratic voters even know Jill Stein?

According to the polls, 12% of respondents are Independents.But the  Gallup Party Affiliation Survey indicates the electorate consists of 42% Independents, 29% Democrats and 29% Republicans.

To believe the Media polls, you must believe that Jill Stein has just 5% of Independents and Democrats. But Bernie had 65-70% of Independents in the primaries- and Jill Stein should be doing nearly as well against Clinton in the polls.

The latest  polls show Trump tied with Clinton and surpassing her in battleground states. The Election Model indicates that he may be leading by 6%. Johnson is taking votes from Trump.  If  Stein’s share increased by 10%,  Clinton’s would decline accordingly – and  Trump would be on his way to a landslide.

Current Media Polls

……………. Pct.. Stein..Clinton.Trump..Johnson

Ind………..12%……5%….40%….40%……..5%

Dem………44%……5%….85%…..5%………5%

Rep……….44%…….0%…..5%….85%…….10%

Total……..100%….2.8%  44.4% 44.4%…. 8.4%

……………. Pct.. Stein..Clinton.Trump..Johnson

Ind………..12%……30%….15%….40%……..15%

Dem………44%……15%….70%…..5%……..10%

Rep……….44%…….0%…..5%….85%……..10%

Total……..100%….10.2%  34.8% 44.4%…. 10.6%

Election Model – Party-ID affiliation from Gallup survey and estimated poll shares.

…………… Pct.. Stein..Clinton.Trump..Johnson

Ind………..42%…35%…20%….25%……20%

Dem ……..29%…20%…70%……5%……..5%

Rep……….29%….2%…..2%…..80% ……16%

Total……..100%..21.1%.29.3% 35.2%….14.4%

http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton_vs_johnson_vs_stein-5952.html

http://www.gallup.com/poll/15370/party-affiliation.aspx

Posted by on September 18, 2016 in 2016 election, Uncategorized

## NH Democratic Primary: another Clinton Miracle?

Richard Charnin
Feb.9, 2016

NH Primary Win Probabilities

It was a Sanders landslide: 60.0-38.4%
http://www.realclearpolitics.com/epolls/2016/president/nh/new_hampshire_democratic_presidential_primary-3351.html#polls

In the final 6 polls Sanders led by 54.5-41.2%. He has an average 99.8% win probability. Combining the polls (2871 sample, MoE = 0.92%) Sanders has a 100% win probability.

In the 2008 NH primary, Obama led the final 7 polls by 38.3-30.0%. His win probability was virtually  100%. Clinton won by 39.0-36.4%, a 10.9% margin shift.

View the polls and probability calculations here.